2010 Buschke Lecture: The Relationship between Local Recurrence and Survival in Breast Cancer Jay R. Harris Dana-Farber Cancer Institute (DFCI) Brigham and Women s Hospital (BWH) Harvard Medical School J. Franz Buschke, M.D. 1902-19831983 Consummate clinician, oncologist, educator, investigator and leader Professor and Director of Radiation Oncology at UCSF from 1957 to 1970 One of the pioneering founders of American Radiation Oncology J. Franz Buschke, M.D. 1902-19831983 The Relationship between Local Recurrence and Survival in Breast Cancer State of the Art in 2000 State of the Art today Gazing forward 1
Progress in Breast Cancer: 2000 By 2000, we had already begun to see a decrease in the mortality rate from breast cancer A key factor in this progress has been the availability of many clinical trials, all of which reside in a common repository in Oxford, UK (EBCTCG) Reasons for Multiplicity of Trials B. Fisher and U. Veronesi In large part, patients willingness to participate in clinical trials Investigators commitment to level I evidence in deciding treatment Early strong commitment of two prominent surgeons to clinical trials 2 surgeons who had the courage in the 1960 s to say they didn t know the answer and that RCT s were needed Bernie Fisher still has the most first author NEJM articles; when he was 75, NEJM gave him a lifetime subscription! 2
Status of Local Therapy in 2000 Status of Local Therapy in 2000 Breast-conserving therapy (BCT) provided alternative to mastectomy Improved breast reconstruction also provided better Quality of Life (QoL) Sentinel node biopsy begins to replace ALND, also improving QoL It was very widely assumed that local therapy impacted local recurrence, but not survival Trials testing variations in local therapy (such as NSABP B-04 and B- 06) failed to show a survival benefit Breast Conserving Therapy (BCT) (My main area of interest) The development of BCT was based on a partnership between surgeons and radiation oncologists, working with pathologists and breast imagers Clinical trials have demonstrated survival equivalent to mastectomy 3
Paradigm Shift! EBCTCG 2005 EBCTCG meta-analysis of trials of local therapy showed a significant and substantial impact of reduced LR on improved long-term survival This survival benefit was achieved either by better surgery or adding RT Ref: EBCTCG, Lancet 366; 2087: 2005 EBCTCG Meta-analysis analysis of Trials of BCS +/- RT NSABP B-06 NSABP B-21 Milan 3 Uppsala-Orebro St. George s Ontario West Midlands CRC UK Swedish Scottish Refs: EBCTCG, Lancet 366; 2087: 2005 Punglia RS et al. NEJM 356; 2399, 2007 Meta-analysis analysis of Trials of BCS +/- RT Oxford Overview This study provided strong evidence that improved local control (surgery or RT) improved survival The individual trials were not large enough ( had insufficient power ) to rule out a 5% survival benefit 4
Oxford Overview Reduction in 5-year LR Reduction in 15-year (not 5-year) mortality No increase in non-breast mortality Similar benefit with post-mastectomy RT and with more surgery A 20% reduction in 5-year LR a 5% reduction in 15-year mortality ( 4:1 ) This Linkage is Strengthened by: Proportionality of the effect: The greater the reduction in 5-year LR, the greater the reduction in 15-year mortality (with a 4:1 ratio) Time course: With > 10% reductions in 5-year LR, the mortality benefit only emerges after 5 years Possible Explanation A hallmark of cancer is genomic instability A recurrent tumor likely has more genetic alterations than the primary For some patients (? 1 in 4), the recurrent tumor has the capacity for metastasis that the primary did not Clinical Implication: Local Therapy is Important! We can no longer be cavalier in our concern about local recurrence Every reasonable measure should be taken to reduce local recurrence We are still trying to determine all the clinical implications of this finding 5
Calculating the Survival Benefit Status of Local Therapy Today A: Estimate the 5-year risk of LR without RT B: Multiply A by < 0.3 to get the risk of 5-year risk of LR with RT C: A B = absolute reduction in LR divided by 4 = estimated reduction in 15-year mortality Local recurrence following BCT has continued to decrease Sentinel node biopsy is very widely used Reconstruction techniques and options have continued to improve Status of Local Therapy Today Data from the EBCTCG and other sources have clearly shown harmful effects if the heart is irradiated RT techniques are available after BCS and after mastectomy to reduce or eliminate cardiac irradiation Cardiac Deaths related to Dose (Ref: EBCTCG, Courtesy Sarah Darby) Estimated Mean Rate Ratio Cardiac Cardiac Dose (Gy) Death for RT/no RT < 5 1.08 (NS) 5-15 1.32 >15 1.63 Risk rate per 10 Gy = 1.31, p < 0.0001 6
Ways to Reduce Cardiac Dose Current Results with BCT Come off midline Cardiac block Prone technique Breath-hold technique Use of separate IMN field (left side) Our results from DFBWCC and MGH are illustrative of the current excellent results seen with BCT These results also illustrate the growing importance of considering biologic subtypes Our Recent Experience 793 BCT patients treated 7/98 12/01 T1 80%, N0 71% Margins: Negative 84%, close 13% ST used in 90% (No Herceptin) Median FU = 70 months 5-year LR = 1.8%! (Ref: Nguyen P et al, JCO 26: 2373, 2008) Subtype is Prognostic for DM Luminal A Favorable Ref: Sorlie, et al PNAS 2003:100, 8418 Luminal B Intermediate Basal and HER2 Unfavorable 7
Is Subtype also Prognostic for LR? Outcome by Biologic Subtype 5-Yr LR 5-Yr DM Subtype approximated by markers: Luminal A = ER or PR+/ HER2- (595) Luminal B = ER or PR+/ HER2+ (77) HER2 = ERPR-/ HER2+ (32) Basal = ER/PR-/HER2- (89) Luminal A 0.8% 3.3% Luminal B 1.5% 12% HER2 8.4% 19% Basal 7.1% 16% On MVA, subtype was only factor significant for LR Similar Results with BCS + RT 5-Yr LR 5-Yr LR* 10-Yr LR^ Luminal A 0.8% 2% 2% Luminal B 1.5% 3% 2% HER2 8.4% 13% 17% Basal 7.1% 21% 16% * NSW, Australia: Millar et al. JCO 27: 4701, 2009 ^ British Columbia: Voduc et al. JCO 28: 1684, 2010 Similar Results in Other Settings 5-Yr LR BCS + RT 5-Yr LR Mast + RT* 5-Yr LR Pre-op, BCS+RT^ Luminal A 0.8% 2% 2% Luminal B 1.5% 3% 2% HER2 8.4% 13% 17% Basal 7.1% 21% 16% * Danish Trials: Kyndi et al. JCO 26: 1419, 2008 ^ MD Anderson (No pcr): Yu et al. SABCS 2009 #957 8
Outcome by Biologic Subtype Reasons for Excellent Outcomes The favorable outcome with Luminal cancers is due to some combination of intrinsic low aggressiveness positive interaction of RT with hormonal therapy Better imaging with mammography (not MRI); use of MRI controversial Better evaluation of the resected breast specimens, especially margins Use of systemic therapy (ST), which greatly improves results of RT 10-Year LR in Recent NSABP Trials (Ref: Wapnir I et al. Proc ASCO 2005) Trial ER 10-Year Status LR (%) B-13 No Chemo - 13.3 B-13 Chemo - 3.5 B-14 No Tamoxifen + 11.0 B-14 Tamoxifen + 3.6 B-19 Chemo - 6.5 B-20 Tam +/- Chemo + 4.7 B-23 Chemo - 4.3 Current Controversies Use of breast MRI at diagnosis Preferential use of mastectomy Use of tamoxifen instead of RT Use of Accelerated Whole Breast RT Use of Accelerated Partial Breast RT Local therapy with preoperative ST 9
Use of Breast MRI at Diagnosis Not Established Decreased Reexcision Decreased LR Improved Survival Established Multicentricity Found Increased Delays Increased Biopsies Increased Costs Increased Mastectomy Effect of Systemic Therapy on LR Systemic therapy by itself reduces LR in the absence of RT This has been evaluated more with hormonal therapy than with chemotherapy There has been interest in the use of tamoxifen instead of RT Can Tamoxifen Substitute for RT?: 5 Published Trials of Tam +/- RT # Patients: Selection FU (med) NSABP B-21 1,009: < 1cm, pn0 87 mths Canadian 769: > 50, T1/2, pn0 67 mths Scottish 427: < 70, T1,2, pn0 67 mths CALGB 636: > 70, T1, c,pn0 95 mths Austrian 869: < 3 cm, gr 1,2, pn0, Tam or AI 54 mths Can Tamoxifen substitute for RT? Tam Tam + RT 5-Yr Endpoint NSABP B-21 8.4% 1.1% LR Canadian 7.7% 13.2% 0.6% 1.1% LR L-RR Scottish 25.0% 3.1% L-RR CALGB 7% 1% Crude L-RR Austrian 5.1% 0.4% LR The results are quite variable and longer FU needed 10
Can Tamoxifen Substitute for RT? Canadian Trial Time to LR Of note, LR increases after 5 years among patients treated with Tam In Canadian and NSABP Trials, LR with Tam is higher at 8 yrs than 5 yrs Raises question whether Tamoxifen is merely delaying LR 7.7% at 5 Yrs 17.6% at 8 Yrs 8-Year LR Rate in NSABP B-21 N=1009 Estimating the Effect of Tam Alone 20 15 10 16.5% 5 9.3% 0 2.3% TAM Radiation Both T1a or T1b, N0 Local Rec Local Rec Estimating is difficult since these trials did not have a no-treatment arm If we assume that RT reduced LR by 70%, then 8-Yr LR in B-21 without any treatment would be 9.3%/.3 = 31% and the reduction with Tam alone would be 47% and when added to RT, it is 75% Fisher B et al. JCO 2002 11
Giving Tamoxifen instead of RT? Tamoxifen seems appropriate in older patients (aged > 70) with ER+ cancers, particularly if serious comorbidities are present Can Chemotherapy Substitute for RT? Few modern studies have administered chemotherapy in absence of breast RT A Canadian trial* of CMFVP (36 wks) vs CMVP-AT (12 wks) included 122 women treated with CS alone; with median FU of 54 mths, ~ 50% of patients had LR In NSABP B-06^, N+ pts treated with CS and chemotherapy had a 12-year cum incidence of LR of 41% *Levine et al. BJC 1992 ^Fisher et al. NEJM 1995 10-Year Results of Canadian Trial of Accelerated WB RT LR Good-Exc Cosmetic Results 50 Gy/25 6.7% 71% 42.5 Gy/16 6.2% 70% In subset analysis, 50 Gy/25 was better for gr 3 cancers Ref: Whelan T et al. NEJM 2010 Can WB RT be Given Faster? Canadian fractionation seems appropriate in older patients (aged > 60) with grade 1, 2 cancers (where a boost has limited value) This has emerged as an alternative to tamoxifen alone 12
Gazing Forward The major focus for progress is further improvements in systemic therapy Local treatment plays a role both in QoL (BCT) and in contributing to the survival rate by decreasing LR Q: What is the role of local treatment with improving systemic therapy? Local Treatment with Improving ST? BCT results will likely get even better with improved ST We can only speculate how improved ST will affect the (4:1) ratio 4:1 was based on the ends of curves derived from older trials without ST Local Treatment with Improving ST? We are just beginning to obtain data on this question We have some data on this from a subset analysis from the Oxford Overview and from the recent retrospective review from the Danish Trial Subset Analysis of PMRT by ST (N+ Patients) ST Used Isolated LR B.C. Mortality Any Death Yes 0.28 0.87 0.88 No 0.30 0.95 0.98 Ref: EBCTCG, Lancet 366; 2087: 2005 (website) 13
Likely Explanation In N+ patients treated with mastectomy and without ST, the correlation of residual local disease and the presence of micrometastatic disease is very high In such patients, reduction of LR is unlikely to improve long-term survival in the absence of ST Danish Trials Results 2 separate PMRT trials: for premenopausal patients, CMF vs CMF + RT and for postmenopausal patients, Tam x 1 year vs Tam + RT Findings in the subset of 1241 patients with ER, PR, HER2 results Ref: Kyndi M et al. JCO 26: 1419, 2008 Danish Trial Results Danish Trial Results 5-Year LR 15-Year Ratio Reduction Mortality Reduction Lum A 22->2%: 20% 11% ~ 2:1 Lum B 39->3%: 36% 23% ~ 1.5:1 HER2 32->13%: 19% - 11% NA Basal 30->21%: 9% 7% ~ 1:1 Limited by retrospective design, small numbers in subgroups, outdated systemic therapy and merging of 2 separate trials Results suggest that Ratio is less than 4:1 with adjuvant therapy and that it varies with subtype/therapy 14
Will the Ratio Stay at 4:1? Local Therapy with Improving ST We don t have enough data to know the answer with certainty The current data suggests that with increasingly effective systemic therapy, the risk of LR will be less, but the ratio will also be less Where we are Where I think we are Ref: Punglia R et al. N Engl J Med 356:2399, 2007 Conclusions Local treatment is important both for QoL (BCT) and for maximizing long-term survival by reducing LR The current estimate is for every 4 LR s avoided at 5 years, there is 1 additional 15-year survivor BCT results have improved, largely due to the interaction of ST and RT Conclusions As ST improves, BCT results will likely get even better As ST improves near term, local treatment will likely become even more important in maximizing survival Eventually, ST will become so good, the role of local treatment will diminish 15
Progress in Breast Cancer It has been a privilege for me to be involved in the breast cancer effort over the past 3 decades I am confident that Dr. Buschke would be very proud of this progress 16